Originally posted by: JPummill, Tue Dec 12, 2006 6:11 pm

Afternoon,

LSF 6.2 is having a problem scheduling jobs in the queue even though there appear to be no reasons why the job(s) should not run. The user has not exceeded his slot quota, there are available slots on the machine, the job will not exceed any assigned parameters, the job submission script is proven to work, etc...

45285 slisenk PEND normal reddiamond dip-gen40 Dec 12 10:56

[root@reddiamond root]# bjobs -l 45285

Job <45285>, Job Name <dip-gen40>, User <slisenk>, Project <default>, Mail <sli
senk@uark.edu>, Status <PEND>, Queue <normal>, Command <#!
/bin/bash; #----------------------------------------------
-------------------------------------------------#;# LSF
settings:
#;#--------------------------
----------------------------------------------------------
-----------#;#;# Name of the job;#BSUB -J dip-gen40;
#;# This option is VITAL for proper initialization of
LSF batch jobs.;#BSUB -L /bin/bash;#;# Output messages
will be written in>
Tue Dec 12 10:56:54: Submitted from host <reddiamond>, CWD <$HOME/Heff_copy/Wor
k/DIP_GEN/Z=40>, Output File <JOB.LOG.%J>, Error File <JOB
.ERROR.%J>, Notify when job begins/ends, 16 Processors Req
uested, Login Shell </bin/bash>;
PENDING REASONS:
New job is waiting for scheduling;

This happened earlier this morning and I was able to get the job(s) to run by doing a "badmin reconfig".

Please assist...

Originally posted by: Rasheed, Thu Dec 14, 2006 3:23 am

Hi,

Please contact support@platform.com for assistance on this issue and our support engineers will be glad to assist you further.

Platform Support

Originally posted by: ddunlap, Thu Dec 28, 2006 8:35 pm

The pending reason for this job is "New job is waiting for scheduling". This means that LSF considers this a "new job".

This is a normal message, when a job is first submitted. Under normal circumstances, the job will be eligible for scheduling within about a minute. At that point, the pending reason will change to something else (or, the job will be scheduled and start running).

If you continue to see the same pending reason, you can run:

bhist -l <jobid>

It may actually be that LSF tried to schedule the job, and for some reason it couldn't, and the job went back to pending. The bhist should show you that.

If the job is just sitting there, and is still considered a "new" job, that indicates some other kind of problem. There are several possibilities, including some of the scheduling parameters in the "lsb.params" file.

My $.02,
Dale
_________________
Dale Dunlap
Technical Consultant